Add an optional timeout for UDP hostport proxy

This commit is contained in:
Tim Hockin 2015-06-29 13:06:02 -07:00
parent 38bddc2a97
commit e54c38d74b
4 changed files with 51 additions and 8 deletions

View File

@ -3,6 +3,7 @@ FROM progrium/busybox
MAINTAINER Tim Hockin "thockin@google.com"
RUN opkg-install socat
ADD start.sh start.sh
# Usage: docker run -p <host-port>:<port> <this-container> <tcp|udp> <port> <service-name>
ENTRYPOINT [ "sh", "-c", "PROTO=$(echo $0 | tr a-z A-Z); exec socat ${PROTO}-LISTEN:$1,reuseaddr,fork ${PROTO}:$2:$1" ]
# Usage: docker run -p <host-port>:<port> <this-container> <tcp|udp> <port> <service-name> [timeout]
ENTRYPOINT ["/start.sh"]

View File

@ -4,7 +4,7 @@
.PHONY: all container push
TAG = v1
TAG = v2
PREFIX = gcr.io/google_containers
NAME = proxy-to-service

View File

@ -7,7 +7,7 @@ also can not choose to expose it on some nodes but not others. These things
will be fixed in the future, but until then, here is a stop-gap measure you can
use.
The container image `gcr.io/google_containers/proxy-to-service:v1` is a very
The container image `gcr.io/google_containers/proxy-to-service:v2` is a very
small container that will do port-forwarding for you. You can use it to
forward a pod port or a host port to a service. Pods can choose any port or
host port, and are not limited in the same way Services are.
@ -23,15 +23,15 @@ metadata:
spec:
containers:
- name: proxy-udp
image: gcr.io/google_containers/proxy-to-service:v1
args: [ "udp", "53", "kube-dns.default" ]
image: gcr.io/google_containers/proxy-to-service:v2
args: [ "udp", "53", "kube-dns.default", "1" ]
ports:
- name: udp
protocol: UDP
containerPort: 53
hostPort: 53
- name: proxy-tcp
image: gcr.io/google_containers/proxy-to-service:v1
image: gcr.io/google_containers/proxy-to-service:v2
args: [ "tcp", "53", "kube-dns.default" ]
ports:
- name: tcp
@ -42,7 +42,14 @@ spec:
This creates a pod with two containers (one for TCP, one for UDP). Each
container receives traffic on a port (53 here) and forwards that traffic to the
kube-dns service. You can run this on as many or as few nodes as you want.
`kube-dns` service. You can run this on as many or as few nodes as you want.
Note that the UDP container has a 4th argument - this is a timeout. Unlike
TCP, UDP does not really have a concept of "connection terminated". If you
need to proxy UDP, you should choose an appropriate timeout. You can specify a
timeout for TCP sessions too, which will close the session after the specified
number of seconds of inactivity. In this case, DNS sessions are not really
ever reused, so a short timeout is appropriate.
[![Analytics](https://kubernetes-site.appspot.com/UA-36037335-10/GitHub/contrib/for-demos/proxy-to-service/README.md?pixel)]()

View File

@ -0,0 +1,35 @@
#!/bin/sh
# Copyright 2015 The Kubernetes Authors All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# Usage: $0 <protocol> <port> <service> [timeout]
# protocol: tcp|udp - case insensitive
# port: port number on which to receive and connect
# service: the destination service name or IP
# timeout: idle timeout in seconds (optional)
if [[ -z "$1" -o -z "$2" -o -z "$3" ]]; then
echo "usage: $0 <protocol> <port> <service> [timeout]"
exit 1
fi
PROTO=$(echo $1 | tr a-z A-Z)
TIMEOUT=""
test -n "$4" && TIMEOUT="-T$4"
CMD="socat ${TIMEOUT} ${PROTO}-LISTEN:$2,reuseaddr,fork ${PROTO}:$3:$2"
echo "Running ${CMD}"
exec ${CMD}